Author Topic: 3458A for less than 1K in 2021 - is it possible? let's find out...  (Read 2962 times)

0 Members and 1 Guest are viewing this topic.

Offline TheSteveTopic starter

  • Supporter
  • ****
  • Posts: 3788
  • Country: ca
  • Living the Dream
I often throw out low offers on ebay, once in a while one gets accepted.

My latest item was a parts/repair 3458A. I was drawn to it for what is a silly reason - it is a very old unit. Serial # 170.
It has all the hallmarks of its age - dimming VFD, aged plastic and cracked/missing bits here and there. Just the way I like'em.

It powered up saying an ACAL ohms was required. It also had a sticker on the front from 2016 saying it was rejected as it failed ohms and current. Under that sticker was the cal sticker from the previous year where it appears to have passed.

At this point we opened the unit up and had a look under the hood. I was expecting to find rev 1 or 2 firmware but was surprised to see factory eproms with rev 9(current) firmware. The NVRAM's had also been changed ~ 2004. When I looked at the bottom of the unit I was greeted with an Agilent branded A3 board - so it had been replaced, likely at the same time the logic board was serviced. The AC board also has an Agilent sticker on it. So they either just added the sticker there or the AC board may have been swapped with a refurb as well. I was a little surprised to see the HI/LOW NVRAM's with a date code 4 years older than the CAL NVRAM. I assume they were all swapped at the same time and the HI/LOW units were just older stock. It looks like all work was done by Keysight(then Agilent).

At this point it was time for a little testing.
Running an ACAL ohms failed with an error indicating cal constant 98 was out of range. Running a normal resistance cal failed on the 10 mAh current range. Cal constant 98 is the stored value for the 10 mAh range which is measured through R211, a 9 ohm resistor on the A1 board. A quick measurement of R211 indicated it had drifted to just over 10 ohms. As a quick test a 100 ohm 1/4 watt resistor was jumpered in parallel to bring it back to ~ 9 ohms. An ohms CAL and ohms ACAL both passed with the quick and dirty fix. Could it be this simple? So far it seems like the resistor was the cause of the rejected cal. The challenge now is to find a properly rated 9 ohm resistor to replace it with. The part is still listed on the Keysight website but is obsolete and they don't have any.

Next was to dump the cal constants via GPIB(not likely useful, but it can't hurt) and then remove the logic board and desolder the NVRAM's. Sockets were added and battery free FM1808 FRAM's were used as replacements. They drop in directly to replace the HI/LOW(U121,U122) rams but an adapter is required to replace the CAL ram.
I also cleaned the unit, thankfully it didn't need to much. Mainly the dust filter and the fan. All of the older stickers were removed and I pushed out a dent that was in the top cover.

For now the 3458A is back together and running with its brothers. It's time for the real gamble to begin - how will the A3 perform. If it doesn't drift horribly the dream of an affordable 3458A is still alive. I'll report back with data in the coming weeks.

And until then does anyone have a suitable 9 ohm resistor recommendation, or perhaps have one from a parts board?

**A special thank you to ManateeMafia who helped get the unit to me, I have no access to my US address due to covid related border closures.**
**While troubleshooting I also did a few google searches for information, all of the results were from xDevs.com - so a shout out to TiN as well!**

Some pics:

Seriously aged LTZ1000A with the early internal LT part number:


Serial # and a curious rams due date:


Rev 9 firmware:


Original NVRAM's, CAL NVRAM 4 years newer for some reason we'll never know:


FRAM adapter and replacements installed:


Cal reject sticker that was on the front:


The quick resistor fix(temporary only of course):

« Last Edit: January 24, 2021, 06:38:43 am by TheSteve »
VE7FM
 
The following users thanked this post: bingo600, jpb, TiN, ManateeMafia, edavid, bck, doktor pyta, wolfy007, bd139, syau

Offline TheSteveTopic starter

  • Supporter
  • ****
  • Posts: 3788
  • Country: ca
  • Living the Dream
Re: 3458A for less than 1K - is it possible? let's find out...
« Reply #1 on: January 24, 2021, 04:42:25 am »
Pics continued:

AC board with Agilent sticker:


A3 board with Agilent branding:


A nice happy picture(so far) with its brothers, the lowest unit is my "reference". The ebay special is on top, value shown after an ACAL, no DCV cal has been performed, let the data logging begin:
« Last Edit: January 24, 2021, 04:49:55 am by TheSteve »
VE7FM
 
The following users thanked this post: TiN, syau

Offline TheSteveTopic starter

  • Supporter
  • ****
  • Posts: 3788
  • Country: ca
  • Living the Dream
Re: 3458A for less than 1K - is it possible? let's find out...
« Reply #2 on: January 24, 2021, 04:59:01 am »
For anyone interested here is the FM1808 FRAM adapter to replace the DS1220 CAL NVRAM. It was designed by volvo_nut_v70 (Ken) and myself. We ordered lots of board so if anyone needs one let me know.

Top/bottom of the blank PCB:


Assembled board - top:


Assembled board - bottom:
« Last Edit: January 24, 2021, 05:11:29 am by TheSteve »
VE7FM
 
The following users thanked this post: doktor pyta, wolfy007, beanflying, iMo, ve2ulu

Offline doktor pyta

  • Frequent Contributor
  • **
  • Posts: 492
  • Country: pl
Re: 3458A for less than 1K in 2021 - is it possible? let's find out...
« Reply #3 on: January 24, 2021, 04:30:08 pm »
Nice try TheSteve!
Now all the forum searches for Micro Precision  ;)

Congrats on the great deal!

Offline Conrad Hoffman

  • Super Contributor
  • ***
  • Posts: 2092
  • Country: us
    • The Messy Basement
Re: 3458A for less than 1K in 2021 - is it possible? let's find out...
« Reply #4 on: January 24, 2021, 05:32:49 pm »
Nice! Just curious about the 9 ohm resistor. How big is it and how good does it have to be? I sometimes make up low value specials by just soldering in the right length of manganin wire. Doesn't work for tiny SMT components but often suited to thru-hole.
 

Offline TheSteveTopic starter

  • Supporter
  • ****
  • Posts: 3788
  • Country: ca
  • Living the Dream
Re: 3458A for less than 1K in 2021 - is it possible? let's find out...
« Reply #5 on: January 24, 2021, 06:39:01 pm »
Now all the forum searches for Micro Precision  ;)

That's just the cal lab.  The RAMS sticker on the back makes me think it was owned by Tektronix though.

I think Keysight themselves add that sticker when the NVRAM board is upgraded/swapped. It is just odd to me that the date appears to be older than the NVRAM's themselves. It is possible a third part installed the rams but would they also have paid the crazy price for factory rev 9 eprom's? It seems like that all of the updates were done late 2005 or 2006.

As you mentioned Micro Precision is a cal lab. The meter also had stickers on it from MACOM and Metelics Corporation(the same company) and it said "CAL LAB" on it. It was sold by a surplus dealer in California. It is likely it was retired when it failed cal in 2016.
VE7FM
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 15234
  • Country: de
Re: 3458A for less than 1K in 2021 - is it possible? let's find out...
« Reply #6 on: January 24, 2021, 10:10:23 pm »
Even if the ADC turn out to be drifty, it still looks like a good deal.

For the 9 Ohms resistor one may get a custom wire would resistor. An alterntive my the some vishey foild resistor. If in a hurry one may settle for 10 Ohms and some 100 Ohms in parallel, just like the current quick fix.  AFAIK the worst case voltage ist at close to 2 V so some 400 mW for the resistor. For the niminal current the power is realtively moderate  (some 1 mW). One should still looke for not to small a form factor to limit self heating and a low TC  (likey < 5 ppm/K). AFAIK TiN had done quite a few tests with replacement resistos. He may even have a used one left over.
 
The following users thanked this post: TheSteve

Offline KK6IL

  • Regular Contributor
  • *
  • Posts: 57
  • Country: us
Re: 3458A for less than 1K in 2021 - is it possible? let's find out...
« Reply #7 on: January 25, 2021, 02:14:34 am »

I've got a spool of Manganin wire, 3.89 ohms/foot if you need a few feet.  We used to use it for trim resistors on a differential multimeter Wavetek tried to make about 1970.  We wound the required amount on the body of a high resistance 1/2 watt carbon resistor and soldered the resistance wire to the resistor leads.  9 ohms might get rather large in diameter and work better on a 1 watt R.
 

Offline martinr33

  • Frequent Contributor
  • **
  • Posts: 363
  • Country: us
Re: 3458A for less than 1K in 2021 - is it possible? let's find out...
« Reply #8 on: January 25, 2021, 04:52:32 am »
That sure is a nice find.

I am not sure that resistor is so fussy, as it sits inside the autocal loop. Texas Components can make a custom resistor for you for a reasonable fee (they made new DC amplifier dividers for me). However, unless the resistor also has a much worse TC than it should, I would say that the parallel trick is fine.

I wonder how much HP used autocal instead of precision resistors to deal with temperature sensitivity. Conversely, had HP used tighter tolerance resistors, would the 3458A need less correction (autocal) over temperature changes?

 

Offline TheSteveTopic starter

  • Supporter
  • ****
  • Posts: 3788
  • Country: ca
  • Living the Dream
Re: 3458A for less than 1K in 2021 - is it possible? let's find out...
« Reply #9 on: January 25, 2021, 05:01:48 am »
The listed specs for the original resistor are - "Resistor-Fixed 9 Ohm +-0.1PCT 0.5W TC+-10 thin film THT", part # 0699-1627.
The Texas Components option is one I am considering.
VE7FM
 

Offline dietert1

  • Super Contributor
  • ***
  • Posts: 2475
  • Country: br
    • CADT Homepage
Re: 3458A for less than 1K in 2021 - is it possible? let's find out...
« Reply #10 on: January 25, 2021, 09:11:11 am »
Another similar resistor is UPW25 or UPW50 at 0.1 % and +/- 3 ppm/K typical. 10R is a stock value and about $ 15. You can combine that one with a 0.1 % MF with +/- 15 ppm/K max. If you want the hassle, select a combination to lower TC.

Regards, Dieter
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 15234
  • Country: de
Re: 3458A for less than 1K in 2021 - is it possible? let's find out...
« Reply #11 on: January 25, 2021, 09:39:25 am »
ACAL does allow to use slightly lower grade resistors. It mainly eliminates the long term drift problem - so it makes it easier to choose a resistor without worrying about the drift, that is hard to test and not often well specified.
The resistors should still be reasonable low TC, as it is not always practical to do an ACAL if the temperature changes.

The large drift of the 9 Ohms resistor is still suspect, so keeping the old one is not a really good option, as chances are it would continue to drift or may fail totally.

Even with ACAL the resistors should be low TC, like <= 10 ppm/K to avoid an effect from self heating and thus non-linearity. For this it helps if the resistor is mechanical not to small so it can get rid of the heat better. Even just 1 mW and a thermal resistance in the 100 K/W range would give 0,1 K temperature rise at 1 mW and thus 1 ppm change in resistance from thermal effects with a 10 ppm/K resistor TC. Luckily a nominal 10 ppm/K resistor is often considerably better than 10 ppm/K TC at a moderate temperature like 30-40 C. Especially with the higher grades, much of the TC is often from the more square parts effecting more the temperature extremes.
So something like UPW50 may work.
 

Offline beanflying

  • Super Contributor
  • ***
  • Posts: 7375
  • Country: au
  • Toys so very many Toys.
Re: 3458A for less than 1K - is it possible? let's find out...
« Reply #12 on: January 25, 2021, 09:39:32 am »
For anyone interested here is the FM1808 FRAM adapter to replace the DS1220 CAL NVRAM. It was designed by volvo_nut_v70 (Ken) and myself. We ordered lots of board so if anyone needs one let me know.



Recent purchaser of a locally sourced 3458A for $1400 USD ($2k AUD equivalent) so they are out there I guess but locally Rocking Horse stuff is more common. Will drop you a PM about the boards as mine has the to be expected issues with advanced old age  :-+
Coffee, Food, R/C and electronics nerd in no particular order. Also CNC wannabe, 3D printer and Laser Cutter Junkie and just don't mention my TEA addiction....
 

Offline MegaVolt

  • Frequent Contributor
  • **
  • Posts: 931
  • Country: by
Re: 3458A for less than 1K - is it possible? let's find out...
« Reply #13 on: January 25, 2021, 09:48:24 am »
Recent purchaser of a locally sourced 3458A for $1400 USD
I envy.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf